Suppressor of cytokine signaling-2 limits intestinal growth and enterotrophic actions of IGF-I in vivo

Carmen Z. Michaylira, James G. Simmons, Nicole M. Ramocki, Brooks P. Scull, Kirk K. McNaughton, C. Randall Fuller, P. Kay Lund

Abstract

Suppressors of cytokine signaling (SOCS) typically limit cytokine receptor signaling via the JAK-STAT pathway. Considerable evidence demonstrates that SOCS2 limits growth hormone (GH) action on body and organ growth. Biochemical evidence that SOCS2 binds to the IGF-I receptor (IGF-IR) supports the novel possibility that SOCS2 limits IGF-I action. The current study tested the hypothesis that SOCS2 normally limits basal or IGF-I-induced intestinal growth and limits IGF-IR signaling in intestinal epithelial cells. Intestinal growth was assessed in mice homozygous for SOCS2 gene deletion (SOCS2 null) and wild-type (WT) littermates at different ages and in response to infused IGF-I or vehicle or EGF and vehicle. The effects of SOCS2 on IGF-IR signaling were examined in ex vivo cultures of SOCS2 null and WT intestine and Caco-2 cells. Compared with WT, SOCS2 null mice showed significantly enhanced small intestine and colon growth, mucosal mass, and crypt cell proliferation and decreases in radiation-induced crypt apoptosis in jejunum. SOCS2 null mice showed significantly greater growth responses to IGF-I in small intestine and colon. IGF-I-stimulated activation of IGF-IR and downstream signaling intermediates were enhanced in the intestine of SOCS2 null mice and were decreased by SOCS2 overexpression in Caco-2 cells. SOCS2 bound directly to the endogenous IGF-IR in Caco-2 cells. The intestine of SOCS2 null mice also showed enhanced growth responses to infused EGF. We conclude that SOCS2 normally limits basal and IGF-I- and EGF-induced intestinal growth in vivo and has novel inhibitory effects on the IGF-IR tyrosine kinase pathway in intestinal epithelial cells.

  • insulin-like growth factor I

the mature intestine is highly responsive to nutritional and physiological status (30). Malnutrition, parenteral feeding, and a number of catabolic diseases can lead to mucosal atrophy. As a response to such physiological or pathophysiological challenges, the intestine is able to adapt by altering the rates of growth, renewal, or functional capabilities of the mucosal epithelium, a phenomenon termed intestinal adaptation (16, 29). Growth factors including IGF-I regulate intestinal adaptation. In animals on total parenteral nutrition (TPN), plasma IGF-I levels are diminished and systemic infusion of IGF-I prevents the mucosal atrophy associated with TPN (1, 3, 8, 39). Systemic administration of IGF-I has been shown to increase intestinal length and mucosal mass in the intact intestine (44) and following bowel resection (2, 18, 24, 31, 55). Transgenic mice overexpressing IGF-I (IGF-I-TG) show increased intestinal length and mucosal mass compared with their wild-type (WT) littermate pairs. In these animal models, IGF-I promotes intestinal growth by increasing crypt cell proliferation and decreasing apoptosis (1, 30, 36, 50).

The trophic effects of IGF-I may be beneficial in situations where adaptive growth is desirable such as TPN or resection. However, considerable evidence also suggests that IGF-I contributes to neoplastic growth of several common cancers, including colorectal cancer (10, 11, 21). Epidemiological studies have shown a positive correlation between circulating levels of IGF-I and the risk of colorectal cancer in healthy individuals (13, 18, 21). Animal studies have also demonstrated an important role of circulating IGF-I levels in colorectal tumor growth. Transplantation of colon adenocarcinoma tissue fragments to the surface of the cecum of liver-specific IGF-I-deficient mice, which show a 25% reduction in serum IGF-I levels, resulted in reduced tumor growth and hepatic metastasis compared with control mice (52). Identifying mechanisms that normally limit growth responses to IGF-I in the intestine would, therefore, be of considerable interest.

The IGF system consists of two structurally related ligands (IGF-I and IGF-II), the IGF-R receptor (IGF-IR), a tyrosine kinase receptor, and six IGF binding proteins (6, 26, 40). IGF-I can regulate growth of the intestine either by endocrine actions of circulating IGF-I or by paracrine/autocrine actions of locally expressed IGF-I (36, 47, 51). Growth hormone (GH) is a major regulator of the circulating and tissue levels of IGF-I (28), stimulating increases in IGF-I. The IGF-IR mediates all of the known biological actions of IGF-I and IGF-II and is expressed in both the small intestine and colon (6). The binding of IGF-I to the IGF-IR leads to autophosphorylation of its tyrosine kinase domain along with phosphorylation of additional tyrosine residues that serve as docking sites for the immediate downstream signaling intermediates, insulin receptor substrate-1 (IRS-1) and Shc, which activate downstream pathways (26). Tyrosine phosphorylation of these proteins leads to the activation of the MAPK and the phosphoinositide 3-kinase (PI3-K) signaling cascades ultimately leading to changes in proliferation and apoptosis (26).

Although intense effort has focused on signaling molecules that mediate IGF-IR action, little is known about mechanisms by which this signaling is terminated or limited in magnitude or duration. SOCS2 belongs to a family of proteins composed of eight members, SOCS-1 to -7 and cytokine inducible SH2-containing protein CIS (19, 20, 32). SOCS proteins are intracellular signaling molecules that limit cytokine receptor signaling via the JAK-STAT pathways (14, 19, 20). Considerable evidence demonstrates that SOCS2 limits the trophic actions of GH. This includes a body overgrowth phenotype in SOCS2 null mice and observations that GH-deficient SOCS2 null mice show enhanced growth responses to exogenous GH (15). Our recent findings that haplotype insufficiency for SOCS2 enhances the effects of a GH transgene on intestinal growth and promotes polyp formation in GH transgenics provide new evidence that SOCS2 limits GH action in the intestine (33). However, as pointed out in a recent editorial, a major unanswered question is whether SOCS2 limits the biological actions of IGF-I, especially in vivo (25). Indirect support for this possibility stems from findings that intestinal overgrowth in GH transgenic mice lacking one functional SOCS2 allele is associated with elevated intestinal IGF-I expression (33). Direct evidence that SOCS2 may inhibit IGF-I action, or signaling via the IGF-IR, is limited to yeast two-hybrid data and data from GST-IGF-IR pull-down assays demonstrating that SOCS2 can bind to the IGF-IR in vitro (4). The current studies therefore used SOCS2 null mice to test specifically whether SOCS2 inhibits basal or IGF-I-induced growth of the intestine. Ex vivo intestinal cultures and cultured intestinal epithelial cells were used to more directly test whether SOCS2 limits IGF-I-mediated activation of the endogenous IGF-IR and downstream signaling pathways. To assess the specificity of results obtained with IGF-I, the effects of SOCS2 deletion on EGF-induced intestinal growth were also determined. Our findings demonstrate that, in the intestine, SOCS2 normally limits the trophic actions of IGF-I and EGF and limits IGF-I signaling in intestinal epithelial cells. To our knowledge, this is the first direct evidence that SOCS2 impacts on the trophic actions of IGF-I in vivo in addition to its documented effects on GH action.

MATERIALS AND METHODS

Animal care and genotyping.

SOCS2 null C57BL/6 mice were provided by Drs. D. Hilton and C. Greenhalgh (Walter and Eliza Hall Institute of Medical Research, Victoria, Australia). SOCS2 heterozygote male and female mice were derived and crossbred to yield SOCS2 null and WT littermates for analyses. Genotyping for all animals was performed on tail DNA by PCR with primers specific for the WT SOCS2 gene (mSOCS2 sense: 5′-CGAGCTCAGTCAAACAGGTAGG-3′; mSOCS2 antisense: 5′-GCTTTCAGATGTAGGGTGCTTCC-3′) or for β-galactosidase (β-gal) sequences present in the disrupted SOCS2 alleles (β-gal sense: 5′-GCAGACGATGGTGCAGGATATCC-3′; β-gal antisense: 5′- GGATCGACAGATTTGATCCAGC-3′). All animal studies were approved by the Institutional Animal Care and Use Committee of the University of North Carolina at Chapel Hill.

Effects of SOCS2 deletion on intestine.

SOCS2 null and WT littermates were studied at 24–28, 45–55, and 100–120 days of age. These times were chosen based on prior observations that the body overgrowth phenotype first occurs at around 42 days of age in SOCS2 null mice (32). Earlier and later time points were chosen as well as the 45- to 55-day time point to assess whether there were age-selective effects of SOCS2 deficiency on intestinal versus body growth. A subcutaneous injection of 5-bromo-2-deoxyuridine (BrdU; 200 mg/kg, Sigma Diagnostic, St. Louis, MO) was administered 90 min before death to permit identification of proliferating crypt cells. After the BrdU injection, mice were anesthetized with pentobarbital sodium (200 μg/g, Abbot Laboratories, Chicago, IL). The entire small intestine and colon were removed, and wet weight and length were recorded. Corresponding segments of jejunum and colon were opened longitudinally and gently scraped with a cold microscope slide to obtain a mucosal enriched fraction and a submucosal/muscularis enriched fraction as previously shown (51). The wet weight of the individual layers per unit length assessed whether the effects of SOCS2 deficiency selectively or preferentially affected the mass of mucosal and muscularis layers. Distal segments of jejunum and colon (0.5 cm) were fixed in 10% formalin and paraffin embedded for morphometry, BrdU localization, and evaluation of apoptosis.

Morphometry.

Morphometry measurements were performed on sections of intestine from 100- to 120-day-old mice because this was the time of maximum effect of SOCS2 deficiency on intestinal weight. Paraffin-embedded, 7-μm sections of jejunum and colon were coded as to genotype of the animal and were stained with hematoxylin and eosin (H&E). Crypt depth, villus height (jejunum only) and thickness of the muscularis propria were measured by a single-blinded observer using light microscopy and computer-assisted morphometry as previously described (51). Six to ten well-oriented villi and a similar number of crypts were measured per segment.

Crypt cell proliferation.

Crypt proliferation was assessed by immunohistochemistry for BrdU using an immunostaining kit (Zymed BrdU staining kit, San Francisco, CA) to label cells in S phase of the cell cycle based on incorporation of BrdU into DNA. Coded sections were scored under a light microscope to assess the number of BrdU-positive cells per crypt. Cells in six to twelve well-oriented crypts were counted per segment.

Apoptosis.

Paraffin-embedded jejunum and colon segments were sectioned at a thickness of 4 μm and H&E stained. Apoptotic cells were identified and counted by a trained individual blinded to the genotype of the sections. Morphological criteria used to identify apoptotic cells included nuclear margination, chromatin and cytoplasmic condensation, shrinkage of neighboring cells, and formation of nuclear bodies as previously described (50). Because spontaneous apoptosis is extremely low, apoptosis was also assessed after γ-irradiation, which induces major increases in apoptosis (50). SOCS2 null and WT mice (100–120 days) received 5 Gy of whole body irradiation delivered at 1 Gy/min with a 137Cs-labeled source as previously described (50). Mice were studied 4 h after irradiation, a time of maximum apoptosis (50). The position of apoptotic cells along the depth of the crypt was recorded with cells at the base of the crypt designated as cell position 1. Relative frequency of apoptosis at different cellular locations in the crypt was assessed as percentage of the total cells at each position from the base of the crypt.

Intestinal responses to infused IGF-I.

Sex-matched WT littermates and sex-matched SOCS2 null littermates (100–120 days) were used to test the effects of IGF-I infusion on intestinal growth. Mice were administered recombinant human IGF-I (rhIGF-I; 2.5 μg·g−1·day−1, Genetech, San Francisco, CA) or vehicle via osmotic minipumps (Alzet microosmotic pumps, 1007D, DURECT, Cupertino, CA) for 5 days. Mice of the same genotype were matched for starting body weight before IGF-I or vehicle infusion. The dose of IGF-I given was at the lowest end of doses used previously to study its enterotrophic actions and was also of the shortest duration. This was done to maximize the possibility of detecting differences in response between SOCS2 null and WT while minimizing the time for compensatory or feedback effects, such as negative feedback suppression of endogenous GH by IGF-I, which might complicate data interpretation. Mice were anesthetized, and pumps were implanted subcutaneously between the scapulae. Mice were monitored daily for changes in body weight. Five days after pump insertion, the mice were killed and small intestine and colon weight and mass were measured. Blood was collected by cardiac puncture, and plasma IGF-I levels were assayed by radioimmunoassay as previously described (35). Growth effects of infused IGF-I were expressed as percent difference in intestinal weight or mass in IGF-I-infused mice compared with vehicle-infused littermates of the same genotype. This allowed us to compare the magnitude of IGF-I action on intestinal growth in SOCS2 null versus WT mice.

Effects of SOCS2 deletion on IGF-IR or STAT activation by IGF-I in intestine ex vivo.

Corresponding segments of small intestine (∼8 cm) were isolated from SOCS2 null and WT mice. The segments were flushed with 1× PBS to remove luminal contents. One-centimeter segments were incubated at room temperature with serum-free medium containing IGF-I (200 ng/ml). After 30–90 min at room temperature, whole extracts were prepared in lysis buffer (50 mM HEPES, 150 mM NaCl, 20 mM Na pyrophosphate, 100 mM NaF, 1.5% Triton X-100, and 10 mM EDTA) containing protease inhibitors (1 ug/ml aprotinin, 1 mM PMSF, and 2 mM vanadate). Extracts were frozen for immunoprecipitation of IGF-IR. Alternatively, for nuclear protein extraction, segments were dounce homogenized in 1 ml ice-cold 1× Tris-buffered saline. Nuclei and nuclear proteins were extracted using standard methods (43). EMSA were performed as previously described (43) on 20 μg of protein using double-stranded oligomers corresponding to a consensus STAT3 (Santa Cruz Biotechnology, Santa Cruz, CA).

Effects of SOCS2 overexpression on IGF-IR activation in intestinal epithelial cells.

Caco-2 cells were obtained from the American Type Culture Collection (Rockville, MD) and propagated in DMEM (GIBCO-BRL, Grand Island, NY) supplemented with 10% FBS, 50 U/ml penicillin, 50 μg/ml streptomycin, and 10 mM HEPES (growth medium). For IGF-I-signaling assays, cells were switched to serum-free medium (SFM) supplemented with 5 μg/ml transferrin and 5 ng/ml selenous acid (Sigma).

The effects of SOCS2 on IGF-I signaling were assessed by infecting Caco-2 cells with an adenovirus expressing FLAG-tagged human SOCS2 (Ad-SOCS2). Ad-SOCS2 was generated by excising the coding sequence for FLAG-tagged human SOCS2 from the pBIG2i-FLAG-SOCS2 expression vector (4) and subcloning into the pShuttle-cytomegalovirus (CMV) shuttling vector using standard protocols. Recombinant, replication-deficient adenovirus was generated from the pShuttle-CMV-SOCS2 construct by the Gene Therapy Center Virus Vector Core at the University of North Carolina at Chapel Hill. An adenovirus expressing green fluorescence protein (GFP) (Ad-GFP, generously provided by the Gene Therapy Center Virus Vector Core, University of North Carolina at Chapel Hill) was used as a control. Caco-2 cells at 60–70% confluence were infected with either Ad-SOCS2 or Ad-GFP using standard protocols. Briefly, growth medium was removed, and a low volume of SFM containing the appropriate virus at 100 multiplicity of infection was added. Twenty-four hours after infection, the adenovirus was washed off and fresh SFM was added. Forty-eight hours after infection, the cells were switched to new SFM plus or minus IGF-I (200 ng/ml). After a 3-min incubation, the cells were harvested in lysis buffer (50 mM HEPES, 150 mM NaCl, 20 mM Na-pyrophosphate, 100 mM NaF, 1.5% Triton X-100, and 10 mM EDTA) containing protease inhibitors (1 ug/ml aprotinin, 1 mM PMSF, and 2 mM vanadate). After centrifugation (12,000 g for 1 min), lysate supernatants were stored at −20°C until further analysis.

Immunoprecipitation and Western immunoblot assays.

Activation of IGF-IR or IRS-1, an early downstream mediator, was assayed in extracts of IGF-I or vehicle-treated intestine or Caco-2 cells by immunoprecipitation of IGF-IR or IRS-I followed by Western immunoblot with an antibody against phosphotyrosine. Total levels of IGF-IR and IRS-1 were also determined by reprobing blots with IGF-IR and IRS-I antibodies. IGF-IR or IRS-1 activation was expressed as the ratio of tyrosine phosphorylated to total IGF-IR or IRS-1. Antiphosphotyrosine (13–5900, Zymed Laboratories, San Francisco, CA), anti-IGF-IR, and anti-IRS-1 (SC-713 and SC-7200, Santa Cruz Biotechnology) antibodies were used for these studies. The ability of SOCS2 to bind to the IGF-IR in infected Caco-2 cells was also tested by immunoprecipitation of the IGF-IR followed by Western immunoblot using an antibody to the FLAG epitope tag (M2 anti-FLAG antibody, Sigma) present in the SOCS2 expression construct. Proteins were visualized with infrared dye-labeled secondary antibodies (IR Dye 800-conjugated anti-goat IgG, Rockland, Gilbertsville, PA and IR Dye 680-conjuated anti-mouse IgG, Molecular Probes, Eugene, OR) and the Odyssey infrared imaging system by LI-COR (Lincoln, NE) according to the manufacturer's instructions. Quantification of the signal was obtained using the Odyssey infrared imaging system application software (Version 1.2, LI-COR).

Intestinal responses to infused EGF.

Intestinal responses to infused EGF were measured to assess whether SOCS2 deficiency may impact on trophic effects of other growth factors that signal via receptor tyrosine kinases. SOCS2 null and WT littermattes were infused with recombinant human EGF (rhEGF; 0.5 mg·g−1·day−1, R&D Systems, Minneapolis MN) or vehicle by minipump, and weight and mass of small intestine and colon were assessed as described for IGF-I.

Statistical analyses.

Values are expressed as means ± SE. One-way ANOVA followed by Fisher's protected least-significant differences multiple t-test was used to test for differences between SOCS2 null and WT or percent change in intestinal growth parameters in IGF-I- or EGF-infused mice versus vehicle infused of the same genotype. Absolute values for plasma IGF-I levels in SOCS2 null and WT were assessed by two-way ANOVA for main effect of genotype or treatment and interactions between genotype and treatment. Statistical significance was set at P < 0.05.

RESULTS

Age-selective effects of SOCS2 deficiency on body and intestinal growth.

Data for body weight and intestinal weight and length are shown on Table 1. As reported previously, SOCS2 null mice first showed increases in body weight compared with WT mice by 45–55 days, and this increase was more pronounced by 100–120 days. Small but significant increases in small intestine length and colon weight were also first observed at 45–55 days. Greater and significant increases in both weight and length of the small intestine and colon were observed in SOCS2 null mice by 100–120 days of age. At this age, effects on small intestinal weight were proportionate to the increases in body weight, whereas the effects on small intestine length and colon weight and length were less than the effects on body weight (Table 1). More detailed growth analysis and subsequent testing of IGF-I action on intestinal growth were performed in 100- to 120-day-old mice because growth effects were greatest at this age.

View this table:
Table 1.

Effects of SOCS2 deletion on body weight and small intestine and colon growth at various ages postweaning

To determine whether SOCS2 deficiency had an effect on intestinal mass or specific effects on mucosal or submucosal/muscularis layers, we calculated intact, mucosal, and submucosal/muscularis mass in jejunum and colon of SOCS2 null and WT mice. We focused on jejunum and colon because we had previously found that levels of SOCS2 expression in these bowel regions in TPN-fed rats inversely correlated with mucosal growth in these segments (34). Compared with WT, SOCS2 null mice had significantly greater intact jejunal mass and mucosal and submucosal/muscularis mass (Fig. 1A). Significant increases in intact and mucosal mass were also observed in the colon (Fig. 1B).

Fig. 1.

Effects of suppressors of cytokine signaling 2 (SOCS2) deficiency on the jejunum and colon segments and mucosal and muscularis masses. A: histograms show wet mass (mg/cm) of intact jejunum and masses of mucosal and submucosal/muscularis layers. B: histograms show wet mass (mg/cm) of intact colon and mass of mucosal and submucosal/muscularis layers. Values are means ± SE; n ≥ 6 wild-type (WT) and SOCS2 null littermate pairs; aP < 0.05 vs. WT.

Morphometry confirmed the effects of SOCS2 deficiency on growth of the jejunum and colon. Villus height, crypt depth, and muscularis thickness were significantly greater in the jejunum of SOCS2 null mice than in WT littermates (Fig. 2). SOCS2 null mice also had significantly greater crypt depth and muscularis thickness in the colon compared with WT mice (Fig. 3). We then assessed whether the increase in mucosal growth was associated with changes in crypt cell proliferation or apoptosis. SOCS2 null mice showed more BrdU-positive cells per crypt compared with WT mice in the jejunum and colon (Fig. 4). Jejunal crypts of SOCS2 null mice showed a small but significant decrease in spontaneous and irradiation-induced apoptosis (Fig. 5A). Interestingly, the decrease in apoptotic cells observed in the jejunum appeared to be localized primarily to cell position 5, counting from the base of the crypt the location of putative stem cells (Fig. 5B). No significant changes in apoptosis were observed in the colon of SOCS2 null and WT mice, although there was a trend for reduced apoptosis in the colon of irradiated SOCS2 null mice vs. WT (Fig. 5C).

Fig. 2.

Effects of SOCS2 deficiency on jejunal crypt-villus axis and muscularis layer. A: representative 4× bright-field microphotographs of hematoxylin and eosin (H&E)-stained jejunum from WT and SOCS2 null mice. B: morphometric measurements of villus height, crypt depth, and muscularis thickness in WT and SOCS2 null jejunum. Histograms show means ± SE; n ≥ 6 WT and SOCS2 null littermate pairs; aP < 0.05 vs. WT.

Fig. 3.

Effects of SOCS2 deficiency on colonic crypt depth and muscularis layer. A: representative 4× bright-field microphotographs of H&E-stained colon from WT and SOCS2 null mice. B: morphometric measurements of crypt depth and muscularis thickness in WT and SOCS2 null colon. Histograms show means ± SE; n ≥ 6 WT and SOCS2 null littermate pairs; aP < 0.05 vs. WT.

Fig. 4.

5-Bromo-2-deoxyuridine (BrdU) incorporation in jejunal and colonic crypt cells. A and B: representative 20× bright-field microphotographs of BrdU-immunostained cells in jejunal and colonic crypts from WT and SOCS2 null mice. C: histograms show mean number of BrdU-positive cells per crypt in jejunum and colon. Values are means ± SE; n ≥ 6 WT and SOCS2 null littermate pairs; aP < 0.05 vs. WT.

Fig. 5.

Decreased apoptosis in SOCS2 null mice. A: histograms show %apoptotic cells per total cells counted in jejunum in normal conditions (nonirradiated) and following irradiation. Values are means ± SE; n ≥ 6; aP < 0.05 vs. WT. B: graph shows the distribution of apoptotic cells in jejunal crypts of SOCS2 null and WT mice after irradiation. The %total apoptotic cells at each position along the length of the crypt is shown. C: histograms show apoptotic cells per total cells counted in colon. Values are means ± SE; n ≥ 6.

Enhanced effects of infused IGF-I on intestinal growth in SOCS2 null mice.

We tested the effects of SOCS2 deficiency on the trophic actions of IGF-I in intestine in vivo by infusing IGF-I or vehicle into age- and sex-matched SOCS2 null or WT mice. The dose and duration of IGF-I administered did not have a significant effect on overall body weight in either SOCS2 null or WT mice, although there was a trend for a small increase in body weight in IGF-I-infused versus vehicle-infused SOCS2 null mice (data not shown). The effects of IGF-I infusion on the intestine were determined as percent difference in weight and mass (mg/cm) of the small intestine and colon in IGF-I-infused versus vehicle-infused mice of the same genotype. In WT mice, this dose and duration of IGF-I infusion had little effect on weight and mass of small intestine or colon, except for a small increase in small intestine weight (Fig. 6A). SOCS2 null mice, however, showed significant increases in weight and mass of small intestine and colon in response to IGF-I, and these increases were significantly greater than the effects in WT mice (Fig. 6, A and B). Because the growth effect of IGF-I was greatest in the colon of SOCS2 null mice, we tested whether IGF-I affected the growth of colonic mucosa and/or muscularis. IGF-I induced parallel increases in the mass of colonic mucosa (20.3 ± 6.7%) and colonic muscularis (22.8 ± 6.9%) in SOCS2 null mice but did not increase mucosal (−3.7 ± 6.7%) or muscularis (−9.2 ± 3.5%) mass in colon of WT mice.

Fig. 6.

Effects of IGF-I infusion on small intestine and colon growth in SOCS2 null and WT mice. A: histograms show %change in small intestine weight and mass (mg/cm) in SOCS2 null and WT mice following 5 days of IGF-I infusion compared with vehicle-infused controls. B: histograms show %change in colon weight and mass in SOCS2 null and WT mice following IGF-I infusion compared with vehicle-infused controls. Values are means ± SE; n ≥ 6 pairs of vehicle or IGF-I-infused animals of each genotype; aP < 0.05 vs. vehicle control; bP < 0.05 in SOC2 null vs. WT.

Plasma IGF-I levels.

Plasma IGF-I levels were measured to verify that IGF-I infusion increased plasma IGF-I. Significant increases in plasma IGF-I were observed in IGF-I-infused WT and SOCS2 null animals (Table 2). Importantly, plasma IGF-I levels were not significantly different in IGF-I-infused or vehicle-infused SOCS2 null compared with WT, verifying that differential intestinal growth responses were not due to differences in plasma IGF-I levels.

View this table:
Table 2.

Plasma IGF-I levels following vehicle or IGF-I infusion

Increased STAT3 activation in intestine of SOCS2 null mice treated with IGF-I.

We first tested the effects of SOCS2 deficiency on STAT3 activation in ex vivo intestinal cultures treated with IGF-I. We focused on STAT3 because of reports in the literature that IGF-I activates STAT3 in other cell types or tissues in vivo (54, 56). As shown in Fig. 7, IGF-I increased nuclear protein binding activity to a STAT3 response element in both WT (1.8 ± 0.3-fold vs. no IGF-I) and SOCS2 null intestine (3.0 ± 0.37-fold vs. no IGF-I). Binding activity was significantly enhanced in SOCS2 null vs. WT intestine.

Fig. 7.

Increased STAT3 activation in intestine of SOCS2 null mice treated with IGF-I. A: autoradiogram of representative EMSA for nuclear binding activity to a STAT3 DNA binding sequence in SOCS2 and WT intestine following a 0-, 30-, 60-, and 90-min treatment with IGF-I. Note the enhanced intensity of the-min time point. cc, Inhibition of shifted complexes by an excess of unlabeled STAT3 oligomer. B: Comassie blue stain of nuclear proteins to show equal amounts of protein in extract from WT and SOCS2 null intestine.

Effects of SOCS2 on IGF-IR activation.

Given the enhanced STAT3 activation by IGF-I in the intestine of SOCS2 null mice, we compared tyrosine phosphorylation of IGF-IR in ex vivo intestinal cultures of small intestine treated with IGF-I. As shown in Fig. 8, across repeated experiments, there was a small but significant increase in tyrosine-phosphorylated IGF-IR expressed as a ratio of total IGF-IR in intestine of SOCS2 null but not WT mice. However, we did not consider this a definitive result because of the small magnitude of the effect. We therefore turned to Caco-2 cells as a simple alternate model system to verify that SOCS2 affects IGF-IR activation in intestinal epithelial cells and to directly determine whether SOCS2 associates with the IGF-IR. Immunoprecipitation studies in Caco-2 cells overexpressing FLAG-tagged SOCS2 or control GFP expression vector revealed that SOCS2 is able to bind to the IGF-IR in the presence or absence of exogenous IGF-I (Fig. 9A). Because previously published biochemical data had suggested that SOCS2 binds only to an activated IGF-IR (4), we examined receptor activation by immunoblotting of the immunoprecipitated IGF-IR with an anti-phosphotyrosine antibody. Constitutive tyrosine phosphorylation of the IGF-IR was observed in Caco-2 cells even in the absence of IGF-I treatment. This is probably due to autocrine secretion and actions of IGFs expressed by these cells (17). Treatment with IGF-I significantly increased levels of tyrosine-phosphorylated IGF-IR in both the Ad-SOCS2- and the Ad-GFP-infected cells, but significantly lower levels of tyrosine-phosphorylated IGF-IR were observed in SOCS2-overexpressing cells compared with GFP-infected control (Fig. 9A). IGF-I-induced tyrosine phosphorylation of IRS-1, a key downstream mediator of IGF-IR signaling (48), was also attenuated in SOCS2 overexpressing cells (Fig. 9B).

Fig. 8.

Tyrosine phosphorylation of IGF-I receptor in intestine of SOCS2 null and WT mice. A: Western immunoblots on immunoprecipitated IGF-IR in whole cell extracts of ex vivo intestinal cultures from WT and SOCS2 null mice after IGF-I treatment. Immunoprecipitates were immunoblotted with antibodies to phosphotyrosine (anti-PY) or IGF-IR. B: histograms show the levels of tyrosine-phosphorylated receptor in IGF-I-treated cultures vs. no treatment controls, expressed as a ratio of total IGF-IR in the same samples. Note the small but significant increase in tyrosine-phosphorylated receptor in intestine of SOCS2 null mice. Values are means ± SE of 3 different experiments; aP < 0.05 vs. no treatment (Tx); bP < 0.05 vs. WT + IGF-I.

Fig. 9.

Binding of SOCS2 to the IGF-I receptor in Caco-2 cells and its effects on signaling. A, left: representative blots of immunoprecipitation (IP) of IGF-IR followed by immunoblotting (IB) for IGF-IR, phosphotyrosine (PY), or FLAG-tagged SOCS2 (FLAG) in Caco-2 cells infected with Ad-SOCS2 or Ad-GFP control and incubated in medium alone (No Tx) or IGF-I. Right: histogram shows fold increase in tyrosine-phosphorylated IGF-I receptor (normalized to total IGF-IR) vs. untreated Ad-GFP control. Values are means ± SE; n = 5; aP < 0.05 vs. Ad-GFP-No Tx; bP < 0.05 vs. Ad-GFP + IGF-I. B, left: representative blots of IP of insulin receptor substrate-1 (IRS-1) followed by IB for IRS-1 or phosphotyrosine (PY) in Caco-2 cells infected with Ad-SOCS2 or Ad-GFP control, and incubated in medium alone (No Tx) or IGF-I. Right: histogram shows fold increase in tyrosine-phosphorylated IRS-1 (normalized to total IRS-1) vs. untreated Ad-GFP control. Values are means ± SE; n = 5; aP < 0.05 vs. Ad-GFP-No Tx; bP < 0.05 vs. Ad-GFP + IGF-I.

Enhanced effects of EGF in intestine of SOCS2 null mice.

To test whether SOCS2 impacts on the trophic effects mediated by other receptor tyrosine kinases, we compared the trophic actions of EGF in SOCS2 null and WT mice. SOCS2 has recently been shown to interact with EGF in regulating neurite outgrowth (12), providing evidence that SOCS proteins may regulate the actions of other growth factors signaling via tyrosine kinase receptors as well as IGF-I. In both jenjunum and colon, EGF significantly increased weight and mass only in SOCS2 null mice and not WT. Differential responsiveness to EGF in SOCS2 null mice was much more dramatic in the colon than the small intestine (Fig. 10).

Fig. 10.

Effects of EGF infusion on small intestine and colon growth in SOCS2 null and WT mice. A: histograms show %change in small intestine weight and mass (mg/cm) in SOCS2 null and WT mice following 5 days of EGF infusion compared with vehicle-infused controls. B: histograms show %change in colon weight and mass in SOCS2 null and WT mice following EGF infusion compared with vehicle-infused controls. Values are means ± SE; n = 5 pairs of vehicle or EGF-infused animals of each genotype; aP < 0.05 vs. vehicle control; bP < 0.05 in SOC-2 null vs. WT.

DISCUSSION

Previous studies in mice homozygous for SOCS2 gene disruption revealed increased body weight and variable increases in organ weight beginning ∼42–56 days of age (15, 32). At this age, SOCS2 deficiency led to increases in carcass, liver, and spleen weight and had little or no effect on the thymus and brain (32); however, the intestine was not examined. Our current observations in the intestine of weanling, 45- to 55-, and 100- to 120-day-old SOCS2 null and WT mice demonstrate that intestinal growth is enhanced in SOCS2-deficient animals, but the effects are minimal until 100–120 days of age and are greater in the small intestine than in the colon. Thus our observations in small intestine and colon provide new evidence for age-selective effects of SOCS2 deficiency on organ growth and for differences in the magnitude of the effect of SOCS2 in different regions of the gastrointestinal tract. To our knowledge, no previous studies have reported age-selective effects of SOCS2 on organ growth. The onset of the body overgrowth phenotype in SOCS2 null mice at 42 days is considerably later than the known time of onset of GH/IGF-I-dependent effects on body growth, which begin around weaning (27), supporting a concept that SOCS2 influences growth preferentially in adulthood. The mechanisms underlying enhanced effects of SOCS2 on body weight or weight of different organs at different times in adulthood is not defined but may reflect the normal timing and levels of SOCS2 expression in different organs during postnatal life.

A key question is whether the increased intestinal growth in SOCS2 null mice reflects enhanced responsiveness of the intestine to IGF-I. Several observations prompted us to hypothesize that SOCS2 affects IGF-I action in the intestine. Crypt proliferation was increased in the small intestine and colon of 100- to 120-day-old SOCS2 null mice. A decrease in spontaneous and irradiation-induced apoptosis was also observed in the jejunum. Multiple models have shown potent effects of IGF-I to increase crypt proliferation and inhibit apoptosis (23, 38). Crypt proliferation is known to increase with age (53), as do levels of IGF-I in the intestine (22), providing indirect evidence that age-selective effects of SOCS2 deficiency on crypt proliferation could reflect increased responsiveness to endogenous IGF-I.

Our findings in IGF-I-infused mice provide direct evidence that SOCS2 deficiency enhances IGF-I action in the intestine. To our knowledge, this is the first direct evidence that SOCS2 normally inhibits the trophic actions of IGF-I in vivo. SOCS2 null mice showed a greater increase in small intestine and colon weight and mass in response to short-term IGF-I infusion compared with WT mice. In these studies, the intestine of WT mice showed little or no growth response to infused IGF-I. The lack of an enterotrophic response in the WT mice after 5 days of low-dose IGF-I treatment was not surprising because previous studies have shown increased mucosal mass with this low dose of IGF-I only after 14 days of treatment (46), and higher doses (12.5 μg/g) have revealed growth effects only after 6.5 days of treatment (30). Had we extended the length of time of IGF-I administration, we would most likely have observed an effect in the WT mice. We deliberately chose a short period of IGF-I administration to avoid or minimize compensatory mechanisms, which could complicate data interpretation. Both colonic mucosa and muscularis showed an increased responsiveness to IGF-I in SOCS2 null mice, suggesting that SOCS2 impacts on IGF-I action in both epithelial and mesenchymal cell-rich layers of the intestine. This is consistent with the ability of IGF-I to increase muscularis as well as mucosal mass in other models (18, 39) and with findings that SOCS2 deficiency affects growth or function of other mesenchyme-rich tissues as well as tissues comprised largely of epithelial cells (32).

Prior evidence that SOCS2 may limit IGF-I action was based on in vitro yeast two-hybrid and GST-IGF-IR pull-down assays indicating that SOCS2 can bind to the IGF-IR (4). To our knowledge, the current studies provide the first evidence that SOCS2 limits the growth-promoting actions of IGF-I in any organ in vivo. Enhanced nuclear STAT3 DNA binding activity and evidence for small but significant increases in tyrosine-phosphorylated IGF-IR in IGF-I-treated intestinal organ cultures provide additional evidence that SOCS2 deficiency alters IGF-I signaling. Findings in Caco-2 cells provide direct evidence that SOCS2 directly binds to the endogenous IGF-IR, limits its activation, and limits activation of the downstream signaling intermediate IRS-I in intestinal epithelial cells. Importantly, this effect was observed in a colon cancer cell line whose growth is highly dependent on IGFs (37, 42) and where we have shown previously that SOCS2 reduces proliferation (34). The mechanisms by which SOCS2 downregulates IGF-I signaling will require further study. SOCS can target some receptors to proteasome-mediated degradation (5), but our current studies revealed no effects of SOCS2 overexpression on levels of IGF-IR, providing indirect evidence arguing against this possibility. SOCS2 contains an SH2 domain and does directly interact with the receptor; thus another possibility is that SOCS2 limits either the tyrosine kinase activity of the IGF-IR or the ability of the tyrosine kinase to autophosphorylate tyrosine residues on the IGF-IR. Defining mechanisms will require further study.

The increased intestinal growth observed with EGF infusion in SOCS2 null mice provides novel evidence to indicate that SOCS2 can limit the actions of another growth factor, which also signals via a tyrosine kinase receptor. Recently, SOCS2 was reported to interact with EGF in mediating neurite outgrowth in cultured neurons and to affect tyrosine phosphorylation of the EGFR (12). Our studies in the intestine provide, to our knowledge, the first evidence that SOCS2 can limit the trophic actions of EGF in intestine.

The effects of SOCS2 deletion to increase crypt proliferation and mucosal mass in the colon of older adult animals, as well as to enhance the trophic actions of IGF-I and EGF on colon, have potential relevance to mechanisms of tumorigenesis in the colon. The colon is a major site of tumor initiation and progression in the intestine. The frequency of colon tumors increases with age, and considerable evidence points to increased circulating IGF-I levels as a risk factor for colon cancer (9). Our findings indicate that SOCS2 normally limits colon growth and crypt proliferation and limits IGF-I and EGF action in the colon of adult animals, suggesting that future studies to test whether SOCS2 could play a tumor suppressor role in the colon are warranted. At present, little information is available about the role of SOCS2 in tumorigenesis, but SOCS2 has been implicated as a tumor suppressor in the hematopoetic system (41), and its expression has been shown to be downregulated in pulmonary adenocarcinoma (49). SOCS2 silencing by promoter hypermethylation has also been observed in ovarian carcinoma cells and in patients with endometrial cancer (7, 45). Our recent studies in SOCS2 null mice crossbred with GH-overexpressing transgenic mice demonstrated that the loss of one SOCS2 allele led to the development of spontaneous but benign lymphoid and hyperplastic polyps in GH-transgenic SOCS2 heterozygotes but not GH transgenics with two functional SOCS2 alleles (33). Given the current data that SOCS2 deficiency amplifies the trophic effects of IGF-I and EGF in the normal colon, it will be of interest to test the effects of SOCS2 in colon cancer models known to depend on or involve EGFR and IGF-IR.

In conclusion, our findings in SOCS2 null mice provide new evidence that endogenous SOCS2 normally limits intestinal growth in older adults but not young animals and that SOCS2 normally limits the trophic actions of IGF-I and EGF on intestine in vivo. In intestinal epithelial cells, SOCS2 limits IGF-I-induced activation of IGF-IR and downstream signaling pathways. Together these studies provide new evidence that SOCS2 directly impacts on the actions of IGF-I, EGF, IGF-IR tyrosine kinase, and STAT3 pathways in the intestine, as well as its documented actions on GH-JAK-STAT pathways in other systems.

GRANTS

This work was supported by National Cancer Institute (NCI) predoctoral research supplement 5 RO1 CA44684–14 (to C. Z. Michaylira), National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-40247 (to P. K. Lund), and NCI predoctoral training Grant CA-72319 (to N. M. Ramocki). The study was facilitated by the molecular histopathology core of the Center for Gastrointestinal Biology and Disease (P30-DK-34987) and the DNA synthesis core of the Lineberger Cancer Center (CA-16086).

Acknowledgments

The authors thank M. Grahn and Dr. D. Ney (University of Wisconsin, Madison, WI) for assistance with plasma IGF-I assays. We also thank Dr. C. Greehalgh for provision of SOCS2 null mice and useful discussions. We acknowledge Dr. R. Furlanetto for provision of SOCS2 expression plasmid. We thank Genentech (South San Francisco, CA) for the generous gift of human recombinant IGF-I.

Footnotes

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

REFERENCES

View Abstract